Google Professional Machine Learning Engineer Question 50
Select 3Google Cloud PlatformYou are a Machine Learning Engineer at a retail company aiming to improve customer support using a conversational AI agent. You decide to use Google Cloud's Agent Builder to create a virtual agent that integrates with your existing customer support system. Which of the following steps should you take to ensure the virtual agent is functional and aligned with your system's needs?
- A
Define intents and training phrases that reflect common customer queries.
- B
Integrate the virtual agent with your customer support system using Dialogflow CX's webhook feature.
- C
Train the virtual agent by providing a large dataset of unstructured text from customer reviews without defining any intents.
- D
Test the virtual agent in a simulated environment and refine responses based on user feedback.
- E
Enable Google Cloud Load Balancer to distribute requests among multiple instances of the agent.
Show answer and explanation
Correct answers: A, B, D
Explanation
When building a virtual agent using Google Cloud's Agent Builder, it is essential to define intents and training phrases to ensure the agent understands user queries. Additionally, integrating the agent with existing backend systems via features like webhooks is necessary for dynamic and context-aware responses. Finally, testing and refining the agent ensures it meets user requirements. Load balancing, while important for scalability, is not a direct step in building or aligning the agent with system needs.
- A. Correct.
Correct: Defining intents and training phrases is a crucial first step in building a virtual agent, as it allows the agent to understand and respond to user queries appropriately.
- B. Correct.
Correct: Webhooks in Dialogflow CX allow integration with backend systems, enabling the agent to fetch data or perform actions necessary for customer support.
- C. Incorrect.
Incorrect: Providing unstructured text without defining intents will not train the virtual agent effectively. Intents and training phrases are necessary to structure the agent's understanding.
- D. Correct.
Correct: Testing the virtual agent in a simulated environment and refining its responses based on feedback ensures its functionality and relevance to user needs.
- E. Incorrect.
Incorrect: While load balancing is important for scaling applications, it is not directly relevant to building or ensuring the functionality of the virtual agent using Agent Builder.